A dashboard is not a decision mechanism
A project dashboard can faithfully report budget, actual cost, committed cost and current forecast while still leaving the weekly operating decision unresolved. The missing question is not “what changed?” but “which feasible intervention should an accountable owner take now, and what outcome is it expected to change?”
That distinction becomes critical when forecast revisions, billing events, change orders and supplier corrections compete for limited PMO and finance capacity.
The cutoff changes the product
A defensible engine reconstructs what was knowable at the decision time. Later outcomes can define labels and measurement, but they cannot leak into the forecast that supposedly preceded them.
The result is a distribution of possible margin and cash outcomes, not a falsely precise point estimate. Policy and contractual constraints then determine which interventions are eligible.
Close the loop on the action
Risk prediction and intervention effect are separate claims. A high probability of margin erosion does not prove that an EAC revision, invoice or change order will protect value. The action must have its own mechanism, evidence, uncertainty, completion event and outcome window.
That is the line between a more attractive report and an operating system that can learn from what actually happened.